When the pandemic hit the world, no one could have predicted the speed in which the semiconductor industry needed to react and make decisions on whether to ramp up or ramp down production.
COVID-19 caused a chain reaction when certain market sectors, such as automotive, forecasted a drop in demand for their products. They reduced their demand for chips accordingly and their freed-up capacity was quickly claimed by other markets that anticipated spikes in demand, such as PCs and consumer electronics.
<<< Start >>>
<<< End >>>
However, the drop in demand wasn’t as significant as expected and the recovery was faster. Capacity was already allocated, which resulted in reporting delays in supply, leaving many companies with chip shortages that halted their entire manufacturing process. The semiconductor industry already has long lead times – from 13 to 18 weeks – and building additional capacity is both time and capital intensive.
Even with building of newer capacity, companies would still have to wait longer than usual times for the chip manufacturing process as well as installation and qualification.
And if the situation could not get any worse, shipments themselves were delayed due to the reduced amount of flights and substrate being in limited supply due to a fire in one of the material’s main factories.
<<< Start >>>
<<< End >>>
Underinvestment in older nodes
This situation also put a spotlight on an underlying problem the industry faced well before the pandemic and subsequent chip shortage: underinvestment in older nodes. Over the past few years, foundries have focused their investments in leading nodes such as 7nm and 5nm because they typically command the best price and are in demand from larger high-tech companies. This has led to relatively lower investment in the older nodes needed by other sectors such as automotive. Thus, when the capacity shortage occurred, the impact on older nodes was even greater and difficult to solve by simply moving production to another fab.
Moreover, doing so could cause significant quality issues because the automotive industry has some of the most stringent requirements in their qualification process. It’s not easy to move chip production to another fab due to the many qualification and certification processes involved and the longer time required to get to a steady state of production.
There is also the well-known impact of quality vs quantity. With semiconductors, quality improves with higher volumes as the greater amount of data generated helps refine the process.
Lessons learned for the future
While little can be done in the immediate term due to capacity being already allocated, there are a few things that companies can do to be prepared when similar issues arise. The first is for businesses to start looking at more manufacturing partners for their products. Another is to be broad and proactive in watching the supply landscape for semiconductors across several industries driving demand fueled by the rise in 5G, autonomous driving, cloud computing, consumer electronics and many more, which are putting more stress on the global semiconductor supply.
Looking at the long-term, the following are some key considerations companies should evaluate to build supply chain resilience into their manufacturing process so that they are better prepared in the future.
Know your supply competitors
Companies typically forecast for their customers based solely on their customers’ needs. While it does make sense, the recent shortage clearly highlighted the need for them to look across the entire ecosystem. When requesting capacity, they should consider which other sectors they might have to compete with –both from a demand and supply perspective.
Traditionally, companies look at competition in terms of who else is selling similar products to their customers. However, supply competitors are equally important to keep in mind. Foundries are serving businesses across different industries and the threat of supply competitors monopolizing capacity might be greater than that of your direct competition.
Having a clear understanding of market trends and expanding one’s definition of competitors will be crucial in deciding how aggressive to be with capacity requirements and avoiding shortages.
<<< Start >>>
Even with building of newer capacity, companies
would still have to wait longer than usual times
for the chip manufacturing
<<< End >>>
Time’s up for just-in-time
Supply chains were built on the just-in-time efficiency principle for all components including
semiconductors, looking to optimize metrics under the assumption that things will always run in sync. However, this recent situation has clearly demonstrated that experiencing semiconductor supply chain disruptions under this principle can have a devastating impact.
For example, when the shortage of a $30 chip can potentially halt the entire production line of a $30,000 car, and with demand for semiconductors ever increasing, it isn’t difficult to see why the threat of chip shortages far outweigh the benefits achieved by just-in-time supply chains.
Every business is a semiconductor business
More industries are using semiconductors and they have become an integral part of the supply chain. As more and more companies look to use semiconductor chips, they may not have the expertise or capabilities internally to understand the landscape, procurement and manufacturing processes. They need to build expertise to better plan and procure chips for their critical products.
Businesses with limited semiconductor expertise in-house should start hiring procurement experts who really understand semiconductors, the supply chain and the sourcing strategy.
This current shortage is a good lesson for any company that relies on chips for their products. It is vitally important to understand that the whole value chain goes six to seven layers deep, and companies need insight into the whole process from the fab, assembly and packaging all the way to the substrate factor - with a view on what’s happening across multiple industries competing for supply.
We are likely to see more and more demand for semiconductors from all sectors of the economy and resulting in continued capacity strain. While there is no silver bullet to solve the current short-term issues, the steps outlined above can help companies avoid the next crisis.